In general, in an image sensor, the number of outputs and an analog digital converter (ADC) resolution are outputs ranging from several M to several tens of M and 10 bits, respectively. On the other hand, in an illuminometer, the number of outputs and the ADC resolution are 1 output and 16 bits, respectively. Thus, when an image sensor is used in the illuminometer, output values are necessary to be added and the dynamic range is necessary to be expanded (e.g. refer to Patent Literature 1).
Therefore, in an image sensor used in an illuminometer (hereinafter, referred to as an image sensor for illuminance), a multiple exposure is generally performed for each frame.
Specifically, the image sensor for illuminance first performs imaging with a longer exposure time and integrates each pixel value of the obtained long-term accumulation image to generate a long-term accumulation value. Then, the image sensor for illuminance performs imaging with a shorter exposure time and integrates each pixel value of the obtained short-term accumulation image to generate a short-term accumulation value. The image sensor for illuminance sets the ratio of the exposure time of a long-term accumulation image to an exposure time of the short-term accumulation image as a gain and multiplies the short-term accumulation value by the gain. The image sensor for illuminance selects, as an illuminance value, one of the short-term accumulation value and the long-term accumulation value obtained by the multiplication by the gain, based on the presence or absence of saturation of accumulated charge in all the pixels at the time of capturing the long-term accumulation image. This allows the dynamic range of the illuminance value to be expanded.
However, when the accumulated charge is saturated only in some pixels at the time of capturing the long-term accumulation image, the linearity of the illuminance values is deteriorated because the long-term accumulation value is selected as the illuminance value. In addition, the short-term accumulation value is multiplied by the gain, and thus when the noise caused by the A/D conversion or the like occurs in the pixel value, the noise of the short-term accumulation value obtained by the multiplication by the gain is larger than that of the long-term accumulation value. Thus, the linearity of the illuminance values is deteriorated when the long-term accumulation value is switched to the short-term accumulation value obtained by the multiplication by the gain.